The following description includes information that may be useful in understanding the present invention. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed invention, or that any publication specifically or implicitly referenced is prior art.
Television audience measurement technologies use human-completed paper logs, somewhat automated “People Meters”, and, more recently, more automated “Portable People Meters” and analysis of “Set-Top Box” data. Paper logs are notebooks in which research subjects record what television broadcast channels and shows they watch and at what time. The paper logs are criticized for being imprecise or inaccurate, for under-reporting daytime and late-night viewing, for failing to record channel “surfing” (rapidly changing channels), and for only measuring audience behavior during relatively few periods during the year. People Meters have buttons, generally one for each research subject in a residence. The research subject presses a button to indicate that they are watching the television and the People Meter records what frequency the television is tuned to. By cross-referencing the time of day with a broadcast schedule for the channel utilizing the frequency, it is possible to estimate the program which the research subject was probably watching (assuming there were no deviations from the schedule).
People Meters also allow non-research subjects to input their age and other demographic information (via buttons), so that non-research subjects may also provide information. Paper logs and People Meters are criticized for requiring active engagement by the research subject, for the selection and distribution of research subjects across the population, for only being used inside of residences, for not measuring audience behavior with respect to non-traditional media rendering devices (smart phones, tables, laptop and desktop computers, and the like), and for the inexact connection between program schedule and what programs and advertisements were actually viewed (program schedules are notoriously unreliable). Portable People Meters (“PPM”) are devices worn on or carried by a research subject. The PPM detects inaudible information encoded in the audio and transmits the decoded information to the research organization. The decoded information identifies media which the research subject was exposed to, at least when the original media was encoded to include the inaudible information.
Set-Top Box data from cable converter boxes and the like has been used more recently to measure audience sizes and characteristics. Set-Top Boxes have a large installed base, the data is easily accessible and there is readily available demographic data at the household level. However, one of the major weaknesses in Set-Top Box data is the inability to verify whether the television screen is actually on and whether the content is being viewed since many people turn off their televisions without turning off the Set-Top Box. This leaves measurement companies guessing and creating algorithms to estimate what was actually viewed. The second issue with Set-top Box data is not knowing definitively what advertisements ran during a program and requires matching of external “as-run ad logs” to determine what ads may have been viewed. This is further complicated by certain advertisement types that are locally inserted, operator inserted, dynamically inserted, or inserted into an “over-the-top” program transmission (program transmission on Netflix, Hulu, and the like is referred to herein as an “over-the-top” or “OTT” transmission). The tracking of advertisements in on-demand programming, OTT programming, and other types of advertisements is virtually impossible via Set-Top Box data.
Previous audience measuring systems are very dependent on the accuracy of a media plan, which is used to determine what the research subject was exposed to; however, anticipated media plans are notorious for being inaccurate relative to what was actually broadcast. Furthermore, existing audience measuring systems are slow, do not record many forums and devices in which and by which media is rendered, and are oriented around shows and show audiences, rather than advertisements and advertisement audiences.
Many “second screen” services exist to provide content on a second screen, such as a smartphone, while a user watches or is present before a first screen. To provide relevant second screen content, such services require knowing what is being rendered on the first screen. Automatic content recognition (“ACR”) is being deployed to automatically recognize content, such as based on recognition of fingerprints or watermarks. However, as the amount of content increases rapidly and as advertisers create more narrowly tailored advertisements and rapidly place them in wide-ranging distribution channels, including in broadcast media (such that advertisement content increases even faster than non-advertisement content), it is not realistic to insert watermarks into all content and fingerprint recognition requires a vast and highly organized infrastructure to characterize the ever-expanding pool of content. As a result, ACR is typically focused on recognizing “shows” in the content, not on recognizing advertisements.
Advertisers attempt to measure the “View Rate” for advertisements (“View Rate” is defined further, herein). However, for broadcast advertisements, the equipment used to measure View Rate is not typically located in the television (or other display device) which renders the advertisement, but is located elsewhere in the path to the television, such as in a Set-Top Box or on a server. Attempting to measure View Rate in the path to the television is problematic, because of disconnects which can occur between the television and the path to the television and because not all users will be connected to a sampled path. However, measuring View Rate in televisions using ACR is prone to the problems noted above. As a result, View Rate measurement, particularly with respect to broadcast media, typically uses small, closely studied, audiences with controlled equipment and/or with limited content access and statistical extrapolation of the resulting information to larger audiences.
Contemporary televisions are capable of reporting a great deal of information, including ACR information (obtained directly or indirectly), however, they do so unreliably. Sometimes the content is misidentified, sometimes a broadcaster or channel is not identified or is misidentified, sometimes the time of day is not reliably identified, sometimes the rendering time is different from the broadcast time (because the content was recorded), sometimes duplicate information is communicated, sometimes there are gaps in what is communicated, sometimes a designated market area (“DMA”) is not reported or is reported unreliably.